GlissEQ vs. Electri-Q

[4damind]

For my ears GEQ and Gliss listens identical. This graph shows a (near identical) HighShelf with both EQs. So its not wrong to say that both playing the same game Can you show us a graph of Electri-Q ?

[Aleksey Vaneev]

Why the 2 creators are inserting posts here in public. I think this damages the opinion on their product (IMHO when a developers try to defend his products it SEEMS he reduces everything to a race).

In my opinion, this is OK to discuss some objective things in public. I'm not arguing on totally subjective things like UI or workflow. But as for the CPU usage, this element can be measured objectively, and if somebody insists GlissEQ uses more CPU resources, this point can be verified.

On Athlon vs Pentium note, relation between efficieny of these two plug-ins should be almost equal when they are compared in pairs on Athlon and on Pentium.

[martian]

Sorry, but I find the GUI of Electri-Q horrible. Electri-Q is no competitor to GlissEQ. But GEQ7 from Kjaerhus and GlissEQ playing on the same high level.

really high level?

electri-q is the same with no oversampling. i think gliss uses something like orfanidis which electri-q has too (turns into shelf), but i dislike that behaviour. just use a proper shelf (or oversampling) if your eq bell doesnt boost the highs the way you want.

Geq with its ACLM technology(?) shows you a symetrical bell on the graph but doesnt deliver, also uses more cpu for seemingly no benefit.

incidently sx q is free (with daw) and has nice symetrical bell, which when set to 20k also lends itself well to a flexible highshelf for an aireq type air boost curve.

without oversampling none of these eqs low shelf like analog cept aireq, but so what? use oversampling if you have it or just use a shelf (or with Q eq a 20k bell). often i prefer the warped lp for a nice deep cut.

[Zaphod (giancarlo)]

In my opinion, this is OK to discuss some objective things in public. I'm not arguing on totally subjective things like UI or workflow. But as for the CPU usage, this element can be measured objectively, and if somebody insists GlissEQ uses more CPU resources, this point can be verified.

On Athlon vs Pentium note, relation between efficieny of these two plug-ins should be almost equal when they are compared in pairs on Athlon and on Pentium.

ok I understand

[martian]

For my ears GEQ and Gliss listens identical. This graph shows a (near identical) HighShelf with both EQs. So its not wrong to say that both playing the same game

if you want symetrical bell gliss also has oversampling, which is apparently good for other reasons too. but i think glisses main purpose is being increasingly lost in this thread, the dynamic mode! there is nothing else like it afaik.

[Zaphod (giancarlo)]

...also refined audiometrics does something similar if I'm not wrong

[Aleksey Vaneev]

martian, next releases of Voxengo equalizers are going to offer an almost ideal bell symmetry - and that's without oversampling (OK, ideal bell symmetry is not possible with biquads at all - due to pure mathematical limits, but can be very close).

[martian]

martian, next releases of Voxengo equalizers are going to offer an almost ideal bell symmetry - and that's without oversampling (OK, ideal bell symmetry is not possible with biquads at all - due to pure mathematical limits, but can be very close).

yes, i saw that on your forum. im not to concerned, since i find i can work around it if need be. but i do prefer symetrical behaviour.

also with the new high/low pass filters, what do you think of this design for a high pass at least, instead of/as well as butterworths (pics in my post): http://www.kvraudio.com/forum/viewtopic ... &start=120

thanks,

[Aleksey Vaneev]

In my opinion, this is OK to discuss some objective things in public. I'm not arguing on totally subjective things like UI or workflow. But as for the CPU usage, this element can be measured objectively, and if somebody insists GlissEQ uses more CPU resources, this point can be verified.

On Athlon vs Pentium note, relation between efficieny of these two plug-ins should be almost equal when they are compared in pairs on Athlon and on Pentium.

ok I understand

Just wanted to correct myself a bit. That 'relation of efficiency' stays the same given both plug-ins use same biquads. There are not so many ways to implement it: with SSE or without, direct form 2 or 1 (these are techie things). If SSE was really as efficient as advertised (it is not on Athlon X2, for example), developers could theoretically decompose 2 serial parametric filters into two parallel filters and thus run equalizer with double efficiency utilizing SSE.

So, it seems my note about 'relation' is not valid in general case: it is valid only in case of simple biquad filters calculated using FPU in series.

[Aleksey Vaneev]

martian, next releases of Voxengo equalizers are going to offer an almost ideal bell symmetry - and that's without oversampling (OK, ideal bell symmetry is not possible with biquads at all - due to pure mathematical limits, but can be very close).

yes, i saw that on your forum. im not to concerned, since i find i can work around it if need be. but i do prefer symetrical behaviour.

also with the new high/low pass filters, what do you think of this design for a high pass at least, instead of/as well as butterworths (pics in my post): http://www.kvraudio.com/forum/viewtopic ... &start=120

thanks,

Group delay plot could tell me a bit more. Here is a group delay plot for Elephant's Bessel high-pass tuned to 40 Hz (log scale, from 20 to 20k Hz) - group delay in samples.

http://www.voxengo.com/temp/bessel_groupdelay_40hz.png
(this pic is in temp folder - I will delete it in the future).

[martian]

Group delay plot could tell me a bit more. Here is a group delay plot for Elephant's Bessel high-pass tuned to 40 Hz (log scale, from 20 to 20k Hz) - group delay in samples.

http://www.voxengo.com/temp/bessel_groupdelay_40hz.png
(this pic is in temp folder - I will delete it in the future).

yes that would seem far more appropriate but christians analyser doesnt like waves ren enough for group delay analysis, and its all i have. can i send you an IR or something?

[Christian Budde]

Group delay plot could tell me a bit more. Here is a group delay plot for Elephant's Bessel high-pass tuned to 40 Hz (log scale, from 20 to 20k Hz) - group delay in samples.

http://www.voxengo.com/temp/bessel_groupdelay_40hz.png
(this pic is in temp folder - I will delete it in the future).

yes that would seem far more appropriate but christians analyser doesnt like waves ren enough for group delay analysis, and its all i have. can i send you an IR or something?

I think it's not that my analyser don't like any plug, but more, that the version you are using, doesn't support displaying the group delay. Try to use the latest version (posted somewhere in the DSP Development forum).

Christian

[martian]

I think it's not that my analyser don't like any plug, but more, that the version you are using, doesn't support displaying the group delay. Try to use the latest version (posted somewhere in the DSP Development forum).

Christian

yes i know the one you gave me has no group delay reading, but the one i was refering to is the one from dsp forum. i get odd squiggles in the group delay readings. basically it feels about right on first measurment, but then when i try to rescale the reading becomes deformed with this square thing.

also if the waveshell is the last plug i used, then the analyser crashes each time i open it, so i have to go into the config settings and change it to something else to load.

[martian]

waves ren, high pass, 60hz, max slope:



gliss, high pass, 24db butter, 60hz:


edit: added gliss comp.

[Christian Budde]

Oh, right I remember, there was still this bug in it. But that doesn't relate to the plugin. It's basically a problem of the analyser. I should upload the bugfixed version...